Automatic Brain Masking

Table of contents

• Description
• Requirements
• Limitations
• Usage
• Setup

Description

Deep learning based project made to automatically mask fetal brains. It can take either individual NIfTI files or the contents of a specified directory.

Currently only a U-net based model is available.

Depending on the input provided (a file or a directory), this tool will look for all .nii files inside of a folder. It will save a new mask with the name name_mask.nii for each .nii file found on the path provided. and it will skip those files that end with mask.nii unless specified otherwise with the --remasking flag.

About this model

This model was trained with a dataset comprising of ~700 reviewed stacks of images and masks, adding to a total of ~20,000 images. It has a wider range of data to improve its robustness. For the sake of identifying this model against priors, its called emerald.

Here are some images showcasing its performance against the previous model, where the $\textcolor{NavyBlue}{\text{blue}}$ mask corresponds to the $\textcolor{NavyBlue}{\text{old}}$ model and the $\textcolor{yellow}{\text{yellow}}$ mask to $\textcolor{yellow}{\text{emeralds}}$.

Requirements

• Python 3
• pip

The following can be installed with the requirements.txt file:

• opencv-python-headless == 4.7.0.68
• MedPy == 0.4.0
• scikit-image == 0.19.3
• keras == 2.11.0
• tensorflow == 2.11.0
• tqdm == 4.64.1
• numpy == 1.24.1

Usage

Its recommended that you create a virtual environment to prevent mixing dependencies. If you don't know how to create one, check out the setup section.

Once you have a virtual environment with all the requirements installed, you can run this tool with the command:

python individual_brain_mask.py  target_file [target_file ...] [-h] [--remasking] [--no-remasking] [--post-processing] \
[--no-post-processing] [--match MATCH [MATCH ...]] [--dilation_footprint SHAPE SIZE] [--no-dilation]

Where:

Argument	Description
target_file	Input path. Required.
-h	Show help message and exit
--remasking	Indicates that images already masked should be remasked, rewritting all *_mask.nii files found. Defaults to false.
--no-remasking	Indicates to skip images that end with *_mask.nii
--post-processing	Indicates that the predicted mask should be post processed (morphological closing and dilation). Defaults to true.
--no-post-processing	Indicates that the predicted mask should not be post processed
--match	Specify if only files with certain words should be masked. Not case sensitive.
--dilation_footprint SHAPE SIZE	Specify the shape and size of the footprint used for dilation. Shapes available are disk and square. If none specified, default is disk 2.
--no-dilation	Masks without dilation.

Limitations

• Unet can currently only work with 256x256 images

Setup

You can create a new virtual environment using the venv command:

python -m venv /path/env_name

This will create a virtual environment called env_name in the directory /path. To activate it, run:

source /path/env_name/bin/activate

The environments name should appear at the beginnig of the shell surrounded by parentheses, like this:

(env_name)$

For further information on how virtual envirionments work, check the python documentation.

Download the tool

Download the source code, cd into your desired location

(env_name)$ git clone https://github.com/sofia-urosa/brain-masking.git
(env_name)$ cd brain-masking

Install requirements from requirements.txt

(env_name)$ pip install -r requirements.txt

And run the tool using the command found in usage.